RuminationMediatestheRelationshipbetween ...downloads.hindawi.com/journals/drt/2011/487873.pdf · birth (formerly named the Wisconsin Maternity Leave and Health Project; [39]). Participants

Hindawi Publishing CorporationDepression Research and TreatmentVolume 2011, Article ID 487873, 9 pagesdoi:10.1155/2011/487873

Research Article

Rumination Mediates the Relationship betweenInfant Temperament and Adolescent Depressive Symptoms

Amy H. Mezulis,1 Heather A. Priess,2 and Janet Shibley Hyde2

1 Department of Clinical Psychology, Seattle Pacific University, 3307 3rd Avenue West Suite 107, Seattle, WA 98119, USA2 Department of Psychology, University of Wisconsin–Madison, 1202 West Johnson Street, Madison, WI 53706, USA

Correspondence should be addressed to Amy H. Mezulis, [email protected]

Received 1 June 2010; Accepted 29 July 2010

Academic Editor: Bettina F. Piko

Copyright © 2011 Amy H. Mezulis et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

This study examined prospective associations between negative emotionality, rumination, and depressive symptoms in acommunity sample of 301 youths (158 females) followed longitudinally from birth to adolescence. Mothers reported on youths’negative emotionality (NE) at age 1, and youths self-reported rumination at age 13 and depressive symptoms at ages 13 and 15.Linear regression analyses indicated that greater NE in infancy was associated with more depressive symptoms at age 15, evenafter controlling for child gender and depressive symptoms at age 13. Moreover, analyses indicated that rumination significantlymediated the association between infancy NE and age 15 depressive symptoms in the full sample. When analyzed separately bygender, however, rumination mediated the relationship between NE and depressive symptoms for girls but not for boys. Theresults confirm and extend previous findings on the association between affective and cognitive vulnerability factors in predictingdepressive symptoms and the gender difference in depression in adolescence, and suggest that clinical interventions designed toreduce negative emotionality may be useful supplements to traditional cognitive interventions for reducing cognitive vulnerabilityto depression.

1. Introduction

Adolescent depression is a major mental health problem.Depression increases in the transition to adolescence, suchthat while fewer than 6% of youth under age 11 willexperience a depressive episode, nearly 20% of youth willexperience a depressive episode by age 18 [1, 2]. In addition,up to 65% of adolescents report subclinical depressive symp-toms at any given time, and extensive research has demon-strated that both mild-to-moderate depressive symptomsand diagnosable depressive episodes predict greater academicand interpersonal problems, substance use, and depressiveepisodes in adulthood [3, 4]. Adolescent depression alsoconfers risk for future depression, with nearly 70% ofadolescents experiencing another episode within five years[5]. Within adolescence, the early to middle adolescentperiod from ages 11 to 15 is of particular salience todepression researchers. During this time, depression ratessurge for all youth and a marked gender difference emerges

such that by age 15 girls are twice as likely as boys to becomedepressed [2].

Extensive research on the etiology of adolescent depres-sion has demonstrated multiple vulnerability factors con-tributing to the rise in depressive symptoms as well as theemergence of the gender difference in depression during thisdevelopmental period. Both affective (e.g., temperament)and cognitive (e.g., rumination, cognitive style) factors havebeen found to confer vulnerability to adolescent depression.In their ABC Model of depression in adolescence, Hyde et al.[6] integrated affective and cognitive models of adolescentdepression by hypothesizing that the specific temperamentaltrait of negative emotionality contributes to the developmentof maladaptive cognitive responses, such as rumination,that become habitual across the adolescent transition andsubsequently confer vulnerability to depression. The currentstudy examines this hypothesis longitudinally in a sampleof community youth followed from infancy into adoles-cence.

2 Depression Research and Treatment

Temperament is conceptualized as biologically-based,relatively stable individual differences in emotional, behav-ioral, and attentional reactivity and regulation [7]. Theseindividual differences are hypothesized to be present earlyin infancy and childhood, and relatively stable across thelifespan. Negative emotionality (NE) is defined as a con-stellation of temperamental characteristics including highfrequency and intensity of negative affective states suchas fear, frustration, and distress. In infancy, children highin NE tend to display strong startle responses to new oraversive stimuli as well as high distress to novel or frustratingsituations. Later in childhood, youth high in NE tend todislike or avoid novel situations, become highly distressed innovel or frustrating situations, and display negative emotionssuch as fear, distress, and frustration more frequently and/orintensely than other children [8, 9]. Numerous studies havedemonstrated an association between NE and depressionamong adults, adolescents, and children [10–13]. Specifi-cally, NE has been demonstrated to be both concurrentlyassociated with and prospectively predictive of depression inadolescence [10, 14].

The cognitive model of depression suggests that indi-vidual differences in cognitive responses to negative eventsmay predispose individuals to becoming depressed whenfaced with such events. One such cognitive vulnerabilityis a ruminative response style [15, 16]. Rumination maybe broadly defined as a passive and perseverative atten-tional focus on negative stimuli, including sad, depressed,or negative emotions, stressful events, and self-critical orotherwise negative thoughts. Nolen-Hoeksema originallydefined rumination broadly as “repetitively focusing on thefact that one is depressed; on one’s symptoms of depression;and on the causes, meanings, and consequences of depressivesymptoms” [15, page 569]. In recent years, researchershave identified many subtypes of rumination differentiatedprimarily by the content upon which an individual is rumi-nating. One such subtype of rumination has been termeddepressive rumination. Depressive rumination is defined asa passive and perseverative focus on negative emotionssuch as sadness and depressed mood [17–20]. Depressiverumination reflects an involuntary coping response in whichone’s attention following a stressor is “directed passively andperseveratively toward the negative emotions elicited by thestressor” [21, page 977]. As such, depressive rumination isessentially emotion-focused and can be differentiated fromother rumination subtypes, including brooding, which isruminative focus on negative or self-critical thoughts [22]and from rumination about other negative emotions such asanger [18, 23].

Rumination on sadness and depressed feelings is believedto prolong and exacerbate depressed mood by increasingthe salience of the negative emotions being attended to. Notsurprisingly, then, depressive rumination has been demon-strated to prospectively predict both the onset and durationof depression among adolescents [24, 25]. However, weunderstand less about the factors contributing to individualdifferences in depressive rumination. Given the emotion-focus of depressive rumination, individual differences innegative emotionality may be associated with individual

differences in the tendency to ruminate on those negativeemotions.

In their ABC Model of depression in adolescence, Hyde etal. [6] integrated affective and cognitive models of depressionby suggesting that youth high in negative emotionalitywould, over time, develop more negative cognitive responsesto stressful events. This hypothesis integrates basic researchlinking affective reactions to stressful events to the cognitiveprocessing of those events, and further frames the inte-gration within the developmental trajectory of adolescentdepression. Weiner [26] noted that affective processing ofnegative events precedes higher-order cognitive processingof the event, such as making attributions about causality.Several studies have demonstrated that affective responsesto stressful events subsequently affect cognitive processing ofthose events. For example, high negative affect is associatedwith greater subjective appraisal of ambiguous or novelevents as stressful [27]. High negative affect is also associatedwith greater interpretation of events as catastrophic, greaterattention to the negative event, greater self-focus, and morenegative expectancies [28–30]. Over time, this pattern ofmore negative and depressogenic cognitive responding maybecome habitual and consolidate into a trait of cognitivevulnerability to depression such as rumination. Severalresearchers have suggested that cognitive vulnerabilitiesto depression emerge and consolidate in early to middleadolescence, a timing that is consistent with the increase indepressive symptoms among youth [31, 32].

This hypothesized developmental link between negativeemotionality and rumination suggests that the predictiverelationship between negative emotionality and depressionmay be mediated in part by rumination. In recent years, ahandful of studies have examined this hypothesis empirically.Feldner et al. reported that, among adults, negative emo-tionality was significantly correlated with rumination [33].A similar correlation has been observed among adolescentsas well [34]. Verstraeten et al. tested the full mediationmodel among adolescents and found that rumination didsignificantly mediate the relationship between negative emo-tionality and depression, but only when constructs wereanalyzed concurrently rather than prospectively [35]. Tothe best of our knowledge, only one study has examinedthe relationship between negative emotionality, rumination,and depression prospectively among adolescents. Mezulis etal. recently found that the prospective relationship betweennegative emotionality in early adolescence (age 12) and laterdepressive symptoms (age 15) was mediated by ruminationat age 14, even after controlling for depressive symptomsat ages 12 and 14 [36]. A limitation of this study, though,is that all measures were self-reported by the youth inadolescence, leaving unanswered the question of whetherindices of temperament early in life are associated with thedevelopment of rumination and subsequent depression inadolescence.

Finally, it is as yet unknown whether prospective relation-ships among negative emotionality, rumination, and depres-sion may contribute to our understanding of the emergentgender difference in depression in adolescence. Althoughmany studies identify a gender difference in rumination that

Depression Research and Treatment 3

has been found to partially mediate the gender differencein depression in adolescence (see [6] for a review), moststudies fail to demonstrate a significant gender differencein negative emotionality in infancy or childhood [37]. It ispossible that negative emotionality may contribute to thedevelopment of rumination amongst girls but not boys,perhaps because of how coping responses to negative moodare socialized differently among boys and girls [38]. Thus,we hypothesized that the prospective relationship betweennegative emotionality and rumination may be stronger forgirls than boys.

The current study extends the extant literature byexamining rumination as a mediator of the relationshipbetween negative emotionality in infancy and depressivesymptoms in adolescence in a prospective study of commu-nity youth followed longitudinally from infancy to age 15.The conceptual model is presented in Figure 1. Specifically,we hypothesized that:

(1) negative emotionality in infancy would predictdepressive symptoms in mid-adolescence (age 15);

(2) the relationship between infant negative emotionalityand adolescent depressive symptoms at age 15 wouldbe mediated by rumination at age 13, even aftercontrolling for depressive symptoms at age 13;

(3) the relationship between infant negative emotionalityand rumination at age 13 would be stronger for girlsthan boys.

2. Method

2.1. Participants. Participants were 301 adolescents (158female) in the United States who have been part of thelongitudinal Wisconsin Study of Families and Work sincebirth (formerly named the Wisconsin Maternity Leave andHealth Project; [39]). Participants were originally recruitedfrom the Madison and Milwaukee, Wisconsin areas andcurrently reside in a range of communities, including alarge Midwestern city, a small Midwestern city, severalsmall towns, and rural areas. Of participants in the presentstudy, 90.0% were White, 4.0% American Indian/Alaskan,3.0% African American, 1.7% Asian/Pacific Islander, 1.0%Hispanic, and 0.3% were members of another group.Participants are ethnically representative of the communitiesfrom which they were recruited, and their families aresocioeconomically similar to families in the United States.

Data were collected at age one and during the summerfollowing grades seven (mean age = 13.52, SD = 0.33;summers of 2004 and 2005) and nine (M = 15.50, SD = 0.33;summers of 2006 and 2007). The present study includesparticipants who completed all measures used in the study.Participants who remained in the study at adolescence didnot differ from those who discontinued participation priorto adolescence in terms of race/ethnicity, family income, orparents’ depressive symptoms.

2.2. Procedure. When the participating children were 12months of age, their mothers completed a written question-naire to assess the children’s temperament. At ages 13 and 15,

Negativeemotionality

(infancy)

Depressivesymptoms

(age 15)

Depressiverumination

(age 13)

Figure 1: Conceptual model.

participants completed a number of questionnaires admin-istered on a laptop computer during in-home visits. Thesequestionnaires included measures of depressive symptoms inthe past two weeks and rumination tendencies.

2.3. Measures

2.3.1. Negative Emotionality. Negative emotionality wasmeasured using the distress to limitations (anger), distressto novelty (fear), and startle subscales of the Infant BehaviorQuestionnaire (IBQ; [40]). The 39 items of these subscalesasked mothers to rate their children’s responses to specificbehaviors (e.g., child became upset when having facewashed) in the past two weeks on a seven-point scale rangingfrom 1 (never) to 7 (always). Negative emotionality wascalculated by averaging the mean scores on each of thethree subscales. Rothbart reported internal consistencies forchildren 12 months of age as .78 for distress to limitationsand .81 for distress to novelty. In the present study, internalconsistency was .84 for distress to limitations, .74 for bothdistress to novelty and startle, and .84 for the combined setof items.

2.3.2. Rumination. Depressive rumination was assessed atage 13 using a short form of the Ruminative Response Scale(RRS) of the Response Style Questionnaire (RSQ; [41]). Theoriginal RRS includes 22 items in which respondents areasked how often they engage in ruminative responses whenthey feel sad or down, with responses rated on a 4-pointLikert scale from 1 (almost never) to 4 (almost always). Inthe current study, we used five items specifically assessingrumination about negative affect. Items include “When Ifeel sad or down, I think about how alone I feel” and“When I feel sad or down, I think about how hard it is toconcentrate”. The five items utilized were selected based uponconsultation with Nolen-Hoeksema at the time of studydesign (personal communication, 2001) as representinga selection of rumination items that excluded automaticnegative thoughts and emphasized instead rumination aboutsad, depressed, or down affect. The exclusion of items thatinclude negative automatic thoughts is preferable becauseit creates a purer measure of depressive rumination that isfocused on the affective component of depressive symptoms,rather than the cognitive component. The full RRS has beenused with adolescents in several prior studies [42, 43]; in thepresent study, internal consistency was .73 for the depressiverumination subscale.


2.3.3. Depressive Symptoms. Depressive symptoms wereassessed at ages 13 and 15 with the Children’s DepressionInventory (CDI; [43]). The CDI includes 27 items assessingcommon affective, behavioral, and cognitive symptoms ofdepression. For each of the 27 items, adolescents were askedto pick which of three statements best described them inthe past two weeks. The three statements represent differinglevels of symptom severity; for example, youth select one ofthe following three statements: “I was sad once in a while,”“I was sad many times,” or “I was sad all the time.” Eachitem is scored 0, 1, or 2, and answers to individual items weresummed, such that a CDI score could range from 0 to 54. TheCDI has demonstrated good internal consistency (typically.71 to .89) and adequate test-retest reliability (typically .72to .87) [44–46]. In the present study, internal consistencywas .83 at age 13 and .86 at age 15.

2.4. Data-Analytic Technique. Analyses were conducted inSPSS as a series of regression models (and in the case ofthe indirect effect, confidence intervals) using the macrocommand set developed by Preacher and Hayes [47] totest mediation models that include covariates. A mediationmodel suggests that the relationship between the predictorvariable (here, negative emotionality) and the outcomevariable (here, depressive symptoms at age 15) is partially orcompletely accounted for by some mediating variable (here,rumination at age 13). With the mediator in the model,the predictor and outcome variables are not expected tobe directly related, but rather indirectly related through theeffect of the predictor variable on the mediator, and then themediator variable on the outcome variable.

Common tests of indirect effects in mediation models,such as the Sobel test, assume that the sampling distributionof an indirect effect is normally distributed; however, thisassumption typically holds only for quite large sample sizes.To avoid violating this assumption, the procedure developedby Preacher and Hayes uses a nonparametric approach thatdoes not require multivariate normality to explicitly testthe indirect effect. Specifically, this procedure employs abootstrap method, in which the original data are sampled(with replacement) 5000 times. The indirect effect coeffi-cients generated from these 5000 samples are then orderednumerically. The low and high values that cap the middle95% of the results represent the bounds of a 95% confidenceinterval. If zero is not within this 95% confidence interval,then there is evidence of an indirect effect between thepredictor variable (e.g., negative emotionality) and outcomevariable (e.g., depressive symptoms) at the standard Type Ierror rate of α = .05.

We examined the hypothesized mediator model threetimes: once for the entire sample, and then separately forboys and girls.

3. Results

3.1. Descriptive Statistics. Table 1 displays descriptive statis-tics for overall negative emotionality, rumination, anddepressive symptoms, separately by gender. Table 2 displays

Table 1: Descriptive statistics for overall negative emotionality,rumination, and depressive symptoms, by gender.

Girls Boys Comparison

Variable M SD M SD t P

NE (infancy) 2.94 .58 2.84 .56 1.46 .146

Rumination (13) 2.02 .55 1.77 .53 3.91 .000

Depression (13) 4.89 5.21 3.79 4.18 2.01 .046

Depression (15) 5.74 5.82 3.84 4.95 3.03 .003

Table 2: Correlation matrix for overall negative emotionality,rumination, depressive symptoms, and gender.

1 2 3 4

(1) NE (infancy) .09 .08 .14

(2) Rumination (13) .17∗ .23∗∗ .15

(3) Depression (13) .10 .52∗∗ .45∗∗

(4) Depression (15) .07 .45∗∗ .62∗∗

Note. Correlations significant at P < .05 denoted by ∗. Correlationssignificant at P < .01 denoted by ∗∗. Correlations are above the diagonal forboys and below the diagonal for girls.

correlations between negative emotionality, rumination, anddepressive symptoms, again separately by gender. As seen inTable 2, the patterns of correlations varied markedly for girlsand boys. Therefore, analyses were computed first for theentire sample and then separately for girls and boys.

3.2. Temperament, Depressive Rumination, and DepressiveSymptoms. The present study examined a mediation modelin which depressive rumination at age 13 was hypothesizedto mediate the relationship between negative emotionalityin infancy and depressive symptoms at age 15. Regressionequations tested each pathway depicted in Figure 1. Inaddition, each equation controlled for prior depressivesymptoms at age 13 and gender. Depressive symptom scoreswere log transformed prior to analysis to account for theirskewed distribution.

Results supported our hypothesized mediation model.The effect of negative emotionality on depressive symptomsat age 15 was significant, b = .09, t(300) = 2.23,P = .03. Additionally, as expected, the path from negativeemotionality to rumination was significant; participantswho were higher in negative emotionality during infancyreported greater tendencies to ruminate as adolescents,b = .14, t(300) = 2.57, P = .01. Similarly, the pathfrom rumination to depressive symptoms was significant;adolescents who reported more rumination at age 13 hadmore depressive symptoms at age 15, even after controllingfor earlier symptoms, b = .12, t(300) = 2.99, P =.003. With rumination in the model, the direct effect ofnegative emotionality on depressive symptoms at age 15was nonsignificant, b = .02, t(300) = .61, P = .54.Additionally, the confidence interval for the effect of theindirect pathway via rumination did not include “0” (.004to.037), indicating a significant mediated pathway. Theoverall mediation model was significant, F(4, 296) = 34.14,P < .001, and accounted for approximately 32% of the


variance in depressive symptoms (R2 = .32, adjusted R2 =.31). Thus, our model of the relationship between negativeemotionality, rumination, and depressive symptoms wassupported by the data.

3.2.1. Temperament, Rumination, and Depressive Symptomsamong Girls. Given the marked differences in the correla-tions among variables for boys and girls in our sample, wealso examined the hypothesized mediation model separatelyby child gender to determine if the relationships amongvariables were comparable across gender. The overall modelfor girls was significant, F(3, 154) = 24.61, P < .001,and explained 32% of the variance in depressive symptoms(R2 = .32, adjusted R2 = .31). The effect of infant negativeemotionality on depressive symptoms at age 15 failed toreach statistical significance, likely as a result of the reducedsample size (b = .08, t(157) = 1.48, P = .14). However,the path from negative emotionality to rumination wassignificant; as expected, girls who were higher in negativeemotionality during infancy reported greater tendencies toruminate as adolescents, b = .16, t(157) = 2.09, P = .04.Similarly, the path from rumination to depressive symptomswas significant; girls who reported more rumination at age13 had more depressive symptoms at age 15, even aftercontrolling for symptoms at age 13, b = .16, t(157) =2.86, P = .005. Finally, the confidence interval for theindirect pathway via rumination did not include “0” (.004 to.061), indicating that rumination significantly mediated therelationship between negative emotionality and depressivesymptoms among girls.

3.2.2. Temperament, Rumination, and Depressive Symptomsamong Boys. The overall model for boys was significant,F(3,139) = 16.04, P < .001, and explained 26% of thevariance in depressive symptoms (R2 = .26, adjusted R2 =.24). The effect of infant negative emotionality on depressivesymptoms at age 15 was comparable to that observed amonggirls and also failed to reach statistical significance, alsolikely as a result of the reduced sample size (b = .08,t(142) = 1.32, P = .19). However, in contrast to girls,the effect of negative emotionality on rumination was notsignificant; boys who were higher in negative emotionalityduring infancy did not report greater tendencies to ruminateas adolescents, b = .09, t(142) = 1.11, P = .27. Furthermore,the effect of rumination on depressive symptoms was also notsignificant; rumination at age 13 did not predict depressivesymptoms at age 15 among boys, b = .08, t(142) = 1.36,P = .18. Finally, the confidence interval for the indirectpathway via rumination did include “0” (−.005 to .030),indicating that rumination did not significantly mediate therelationship between negative emotionality and depressivesymptoms among boys.

4. Discussion

This study examined the relationship between temperament,rumination, and depressive symptoms prospectively in acommunity sample of youth followed from infancy to age

15. The primary purpose of the study was to empiricallytest one of the integrative hypotheses of the ABC Modelof adolescent depression which asserts that affective vulner-ability to depression, that is, temperament, contributes tothe development of cognitive vulnerability to depression,that is, rumination [6]. This hypothesis suggests that indi-viduals who are temperamentally predisposed to respondto stressful events with intense and prolonged negativeaffect will subsequently allocate more attentional resourcesto those events, and that this pattern of affective-cognitiveprocessing of stressful events will, over time, consolidateinto the stable cognitive vulnerability of rumination. Thepresent study extends prior examinations of rumination asa mediator of the relationship between temperament anddepression by examining infant temperament as it predictslater rumination and depressive symptoms.

4.1. Infant Negative Emotionality and Adolescent Depres-sion. Numerous studies have examined the temperamentalconstruct of negative emotionality as it predicts depressivesymptoms and disorders [7, 13]. However, only a handful ofstudies have examined the predictive relationship betweeninfant or early childhood temperament on depression inadolescence. Moffitt et al. [48] reported that behavioralinhibition, one component of negative emotionality, at age3 prospectively predicted depression diagnoses by age 21.Similarly, Lonigan et al. found that childhood negativeemotionality at age 9 predicted depressive symptoms at age16 [49]. Consistent with these prior studies, we similarlyfound that greater negative emotionality in infancy wassignificantly associated with greater depressive symptoms inadolescence.

4.2. Linking Affective and Cognitive Processes to AdolescentDepression. The primary purpose of this study, however,was to examine the mechanism by which early individ-ual differences in negative emotionality develop into laterdepressive symptoms. In adolescence, we know that indi-vidual differences in affective and cognitive responses tostressful events may differentiate individuals for whom mooddisturbances are transient from individuals for whom thatmood disturbance persists and develops into a depressiveresponse. Teasdale’s differential activation hypothesis [30]suggests that some individuals are more likely to respondto negative affect with the activation of negative thoughtsand rumination. Teasdale labeled this individual differencein the extent to which negative cognitive processing iselicited by negative affect as cognitive reactivity, and suggestedthat for these cognitively reactive individuals, a viciouscycle between negative affect and negative thinking willensue that eventually leads to depression. The ABC Modelpresents a theoretically consistent hypothesis that emphasizesthe developmental relationship between negative affect andcognitive processing, suggesting that it is early individualdifferences in affective responding to stress that set thedevelopmental stage for this vicious cycle of affective andcognitive processing to ensue. We hypothesized that childrenwho are temperamentally high in negative emotionality will


become adolescents who are high in cognitive vulnerability,including rumination.

Only a handful of studies have examined whether indi-vidual differences in negative emotionality predict individualdifferences in cognitive vulnerability to depression. Severalstudies have examined the relationship between neuroticism,rumination, and depression among adults [50–55]. Neuroti-cism is a personality trait which is similar to, but more broadthan, the temperamental construct of negative emotionality;neuroticism is associated with high negative emotionality aswell as high stress sensitivity and worry [56].

Few studies have examined the relationship between thespecific construct of negative emotionality and rumination.Our findings contribute to a small but growing bodyof literature suggesting that negative emotionality is animportant contributor to rumination among adolescents.Our findings are consistent with those of Chang [34] andVerstraeten et al. [35] demonstrating a significant associationbetween these constructs, and providing further evidencethat the relationship between temperament and depressionmay be mediated by cognitive processes.

Interestingly, although the mediation model was sup-ported for the entire sample, follow-up analyses by child gen-der suggested that the prospective association between infantnegative emotionality and adolescent depressive symptomsmay be mediated by rumination among girls but not amongboys. This finding suggests, consistent with the ABC Modelof the gender difference in adolescence depression, thatthere may be multiple processes contributing to the genderdifference in depression. In our sample, girls and boys didnot differ on mother-reported infant negative emotionality,a finding consistent with a recent meta-analysis examininggender differences in temperament [37]. However, boys andgirls did differ on rumination in early adolescence, withgirls reporting greater rumination than boys. Our findingssuggest that early negative emotionality may contribute tolater ruminative tendencies among girls but not boys. Itis interesting to speculate on this gender divergence; onepossible explanation may be how parents respond to displaysof negative emotionality differently for sons and daughters.Recent research has suggested that mothers may be morelikely to direct attention to and encourage discussion ofnegative emotions, particularly fear, distress, and sadness,when interacting with their daughters than with their sons,and that this gender difference in parenting style maycontribute to the gender difference in rumination [38].

It is also important to note that negative emotionality isonly one of multiple constructs of temperament associatedwith vulnerability to depression. Prior research has alsoimplicated low positive emotionality in the etiology ofdepression [12], although that relationship is not hypothe-sized to be mediated by cognitive vulnerability [6]. Recentresearch has also suggested that the relationship betweennegative emotionality and cognitive vulnerability may itselfby moderated by other regulatory components of temper-ament. For example, effortful control is another featureof temperament conceptualized as “the ability to inhibit adominant response to perform a subdominant response”[7, page 137], and thus is a self-regulatory process that

requires effortful or voluntary control of both attention andbehavior to modulate emotional experience and expression[57]. Some have found that effortful control may moderatethe relationship between the affective reactivity of negativeemotionality and rumination [35]. Although the presentstudy did not include a measure of effortful control withwhich to examine this hypothesis, it would be an interestingelaboration for future prospective studies.

4.3. Clinical Implications. Adolescent-onset depression isassociated with both concurrent deficits in adaptive func-tioning and prospective risk for future depressive episodes.The vast majority of youth experiencing depression inadolescence will have another episode within five years[5]. Unfortunately, treatments for adolescent depression lagbehind those for other disorders and more than half of youthfail to respond to currently available interventions [58].The majority of current depression interventions emphasizetechniques designed to reduce or eliminate depressogeniccognitive processes such as rumination. Continued evidencesuggesting that individual differences in negative emotion-ality significantly contribute to individual differences inrumination suggest that interventions designed to reduceindividuals’ negative affect may be helpful in treating orpreventing depression as well. Relaxation training maybe effective in attenuating youths’ automatic and intensenegative affective responses, and a growing body of researchsuggests that mindfulness-based interventions may improveemotion regulation through improving the individual’sability to respond to stressful situations reflectively ratherthan automatically. Mindfulness has been shown to beeffective in treating depression, possibly through its effectsin decreasing rumination [59, 60]. Kabat-Zinn [61] alsosuggested that the slow and deep breathing taught inmindfulness training may reduce physiological reactivityand the subjective emotional and physiological feelings ofdistress. Although few studies have examined the effectsof mindfulness on affective responding, Goldin and Gross[62] recently reported that individuals with social anxi-ety reported less negative affect during a breath-focusedmindfulness task compared to a distraction-focused task.Similarly, Charbonneau and Mezulis [63] recently reportedthat among college females high in negative emotionality,a brief intervention teaching emotion regulation strategiessuch as deep breathing and progressive muscle relaxationstrategy was effective in reducing rumination. In summary,a greater understanding of the relationship between affectiveand cognitive vulnerability to depression may suggest thatclinical interventions targeting negative emotionality oremotional reactivity may be effective in reducing cognitivevulnerability to depression as well.

4.4. Limitations and Future Research. While our studydemonstrates links between temperament, rumination, anddepressive symptoms, several limitations should be noted.First, the study examined rumination narrowly defined asperseverative attention on negative emotions. Conceptually,the depressive rumination subtype of rumination is mostlogically linked with negative emotionality, as the strong


negative emotions experienced by individuals high in thistemperamental feature may particularly elicit the attentionalfocus on negative emotions that comprises depressive rumi-nation. However, future research may want to consider therelationship between negative emotionality and other formsof rumination. Of particular interest may be brooding, whichis defined as rumination on negative or self-critical thoughts.Recent evidence specifically links the rumination dimensionof brooding with depression among adolescents and may beone specific facet of the cognitive reactions to stress elicitedby negative emotionality [64, 65]. Second, our sample wasa community sample of predominantly Caucasian youth.The relationships among our constructs may be differentamong high-risk or clinical samples. Finally, we note that thepathway from negative emotionality to depression, mediatedby rumination, demonstrated in the current sample is butone of multiple developmental trajectories implicated in theetiology of adolescent depression. As extensively reviewedin other studies, there are multiple temperamental andcognitive factors that contribute to adolescent depressionboth within and across individuals. Low positive emotion-ality is another temperamental factor implicated in theetiology of depression, as well as other cognitive vulnerabilityfactors such as negative cognitive style (as reviewed in [6]).Thus, the negative emotionality-rumination-depression linkexamined here is undoubtedly but one pathway to adolescentdepression.

Despite these limitations, our findings continue toimplicate temperament in the development of cognitivevulnerability to depression. These findings contribute toour ability to identify at-risk individuals as well as designinterventions targeting both affective and cognitive processesin adolescent depression.

Acknowledgment

This research was supported, in part, by Award no.F31MH084476 from the National Institute of Mental Healthto H. A. Priess. The content is solely the responsibility of theauthors and does not necessarily represent the official viewsof the National Institute of Mental Health or the NationalInstitutes of Health.

References

[1] P. Cohen, J. Cohen, S. Kasen et al., “An epidemiological studyof disorders in late childhood and adolescence I. Age- andgender-specific prevalence,” Journal of Child Psychology andPsychiatry, vol. 34, no. 6, pp. 851–867, 1993.

[2] R. C. Kessler, K. A. McGonagle, M. Swartz, D. G. Blazer, and C.B. Nelson, “Sex and depression in the National ComorbiditySurvey I: lifetime prevalence, chronicity and recurrence,”Journal of Affective Disorders, vol. 29, no. 2-3, pp. 85–96, 1993.

[3] D. M. Fergusson and L. J. Woodward, “Mental health,educational, and social role outcomes of adolescents withdepression,” Archives of General Psychiatry, vol. 59, no. 3, pp.225–231, 2002.

[4] I. H. Gotlib, P. M. Lewinsohn, and J. R. Seeley, “Consequencesof depression during adolescence: marital status and marital

functioning in early adulthood,” Journal of Abnormal Psychol-ogy, vol. 107, no. 4, pp. 686–690, 1998.

[5] P. M. Lewinsohn, G. N. Clarke, J. R. Seeley, and P. Rohde,“Major depression in community adolescents: age at onset,episode duration, and time to recurrence,” Journal of theAmerican Academy of Child and Adolescent Psychiatry, vol. 33,no. 6, pp. 809–818, 1994.

[6] J. S. Hyde, A. H. Mezulis, and L. Y. Abramson, “The ABCsof depression: integrating affective, biological and cognitivemodels to explain the emergence of the gender difference indepression,” Psychological Review, vol. 115, no. 2, pp. 291–313,2008.

[7] M. Rothbart and J. Bates, “Temperament,” in Handbook ofChild Psychology, W. Damon, Ed., vol. 3, pp. 105–176, JohnWiley & Sons, New York, NY, USA, 1998.

[8] J. Belsky, K.-H. Hsieh, and K. Crnic, “Infant positive andnegative emotionality: one dimension or two?” DevelopmentalPsychology, vol. 32, no. 2, pp. 289–298, 1996.

[9] A. Buss and R. Plomin, “The EAS approach to temperament,”in The Study of Temperament: Changes, Continuities, andChallenges, R. Plomin and J. Dunn, Eds., Erlbaum, Hillsdale,NJ, USA, 1986.

[10] J. L. Anthony, C. J. Lonigan, E. S. Hooe, and B. M. Phillips,“An affect-based, hierarchical model of temperament andits relations with internalizing symptomatology,” Journal ofClinical Child and Adolescent Psychology, vol. 31, no. 4, pp.480–490, 2002.

[11] C. R. Colder, J. A. Mott, and A. S. Berman, “The interactiveeffects of infant activity level and fear on growth trajectoriesof early childhood behavior problems,” Development andPsychopathology, vol. 14, no. 1, pp. 1–23, 2002.

[12] I. M. Goodyer, L. Ashby, P. M. E. Altham, C. Vize, and P. J.Cooper, “Temperament and major depression in 11 to 16 yearolds,” Journal of Child Psychology and Psychiatry, vol. 34, no. 8,pp. 1409–1423, 1993.

[13] S. B. Roberts and K. S. Kendler, “Neuroticism and self-esteemas indices of the vulnerability to major depression in women,”Psychological Medicine, vol. 29, no. 5, pp. 1101–1109, 1999.

[14] B. E. Compas, J. Connor-Smith, and S. S. Jaser, “Tempera-ment, stress-reactivity, and coping: implications for depres-sion in childhood and adolescence,” Journal of Clinical Childand Adolescent Psychology, vol. 33, no. 1, pp. 21–31, 2004.

[15] S. Nolen-Hoeksema, “Responses to depression and theireffects on the duration of depressive episodes,” Journal ofAbnormal Psychology, vol. 100, no. 4, pp. 569–582, 1991.

[16] S. Nolen-Hoeksema and R. Plomin, “The response style the-ory,” in Depressive Rumination: Nature, Theory and Treatment,C. Papageorgiou and A. Wells, Eds., pp. 107–123, John Wiley& Sons, West Sussex, UK, 2004.

[17] M. Conway, P. A. R. Csank, S. L. Holm, and C. K. Blake, “Onassessing individual differences in rumination on sadness,”Journal of Personality Assessment, vol. 75, no. 3, pp. 404–425,2000.

[18] M. Peled and M. M. Moretti, “Rumination on anger andsadness in adolescence: fueling of fury and deepening ofdespair,” Journal of Clinical Child and Adolescent Psychology,vol. 36, no. 1, pp. 66–75, 2007.

[19] F. Raes, D. Hermans, J. M. G. Williams, P. Bijttebier, and P.Eelen, “A “triple W”-model of rumination on sadness: why amI feeling sad, what’s the meaning of my sadness, and wish Icould stop thinking about my sadness (but I can’t!),” CognitiveTherapy and Research, vol. 32, no. 4, pp. 526–541, 2008.

[20] L. Rood, J. Roelofs, S. M. Bogels, S. Nolen-Hoeksema, and E.Schouten, “The influence of emotion-focused rumination and


distraction on depressive symptoms in non-clinical youth: ameta-analytic review,” Clinical Psychology Review, vol. 29, no.7, pp. 607–616, 2009.

[21] J. K. Connor-Smith, B. E. Compas, M. E. Wadsworth, A.H. Thomsen, and H. Saltzman, “Responses to stress inadolescence: measurement of coping and involuntary stressresponses,” Journal of Consulting and Clinical Psychology, vol.68, no. 6, pp. 976–992, 2000.

[22] W. Treynor, R. Gonzalez, and S. Nolen-Hoeksema, “Rumina-tion reconsidered: a psychometric analysis,” Cognitive Therapyand Research, vol. 27, no. 3, pp. 247–259, 2003.

[23] W. Gerin, K. W. Davidson, N. J. S. Christenfeld, T. Goyal, andJ. E. Schwartz, “The role of angry rumination and distractionin blood pressure recovery from emotional arousal,” Psychoso-matic Medicine, vol. 68, no. 1, pp. 64–72, 2006.

[24] R. J. Park, I. M. Goodyer, and J. D. Teasdale, “Effects of inducedrumination and distraction on mood and overgeneral auto-biographical memory in adolescent major depressive disorderand controls,” Journal of Child Psychology and Psychiatry, vol.45, no. 5, pp. 996–1006, 2004.

[25] J. S. Silk, L. Steinberg, and A. S. Morris, “Adolescents’ emotionregulation in daily life: links to depressive symptoms andproblem behavior,” Child Development, vol. 74, no. 6, pp.1869–1880, 2003.

[26] B. Weiner, “An attributional theory of achievement motivationand emotion,” Psychological Review, vol. 92, no. 4, pp. 548–573, 1985.

[27] P. T. Costa, M. R. Somerfield, and R. R. McCrae, “Personalityand coping: a reconceptualization,” in Handbook of Coping:Theory, Research, Applications, M. Zeidner and N. S. Endler,Eds., John Wiley & Sons, Oxford, UK, 1996.

[28] D. Derryberry and M. A. Reed, “Temperament and attention:orienting toward and away from positive and negative signals,”Journal of Personality and Social Psychology, vol. 66, no. 6, pp.1128–1139, 1994.

[29] T. Pyszczynski, K. Holt, and J. Greenberg, “Depression, self-focused attention, and expectancies for positive and negativefuture life events for self and others,” Journal of Personality andSocial Psychology, vol. 52, no. 5, pp. 994–1001, 1987.

[30] J. D. Teasdale, “Cognitive vulnerability to persistent depres-sion,” Cognition and Emotion, vol. 2, pp. 247–274, 1988.

[31] J. R. Z. Abela and B. L. Hankin, “Cognitive vulnerabilityto depression in children and adolescents: a developmentalpsychopathology approach,” in Handbook of Depression inChildren and Adolescents, J. R. Z. Abela and B. L. Hankin, Eds.,pp. 35–78, Guilford Press, New York, NY, USA, 2007.

[32] A. H. Mezulis, J. S. Hyde, and L. Y. Abramson, “Thedevelopmental origins of cognitive vulnerability to depression:temperament, parenting, and negative life events in childhoodas contributors to negative cognitive style,” DevelopmentalPsychology, vol. 42, no. 6, pp. 1012–1025, 2006.

[33] M. T. Feldner, E.W. Leen-Feldner, M. J. Zvolensky, and C.W. Lejuez, “Examining the association between rumination,negative affectivity, and negative affect induced by a pacedauditory serial addition task,” Journal of Behavior Therapy andExperimental Psychiatry, vol. 37, no. 3, pp. 171–187, 2006.

[34] E. C. Chang, “Distinguishing between ruminative and distrac-tive responses in dysphoric college students: does indication ofpast depression make a difference?” Personality and IndividualDifferences, vol. 36, no. 4, pp. 845–855, 2004.

[35] K. Verstraeten, M. W. Vasey, F. Raes, and P. Bijttebier,“Temperament and risk for depressive symptoms in adoles-cence: mediation by rumination and moderation by effortful

control,” Journal of Abnormal Child Psychology, vol. 37, no. 3,pp. 349–361, 2009.

[36] A. Mezulis, J. Simonson, E. McCauley, and A. Vander Stoep,“Brooding, not reflection, mediates the association betweentemperament and depressive symptoms in adolescence,”under review.

[37] N. M. Else-Quest, J. S. Hyde, H. H. Goldsmith, and C. A. VanHulle, “Gender differences in temperament: a meta-analysis,”Psychological Bulletin, vol. 132, no. 1, pp. 33–72, 2006.

[38] S. J. Cox, A. H. Mezulis, and J. S. Hyde, “The influenceof child gender role and maternal feedback to child stresson the emergence of the gender difference in rumination inadolescence,” Developmental Psychology, vol. 46, no. 4, pp.842–852, 2010.

[39] J. S. Hyde, M. H. Klein, M. J. Essex, and R. Clark, “Maternityleave and women’s mental health,” Psychology of WomenQuarterly, vol. 19, no. 2, pp. 257–285, 1995.

[40] M. K. Rothbart, “Measurement of temperament in infancy,”Child Development, vol. 52, pp. 569–578, 1981.

[41] S. Nolen-Hoeksema and J. Morrow, “A prospective studyof depression and posttraumatic stress symptoms after anatural disaster: the 1989 Loma Prieta earthquake,” Journal ofPersonality and Social Psychology, vol. 61, no. 1, pp. 115–121,1991.

[42] P. C. Broderick and C. Korteland, “A prospective study ofrumination and depression in early adolescence,” ClinicalChild Psychology and Psychiatry, vol. 9, no. 3, pp. 383–394,2004.

[43] P. E. Jose and I. Brown, “When does the gender difference inrumination begin? Gender and age differences in the use ofrumination by adolescents,” Journal of Youth and Adolescence,vol. 37, no. 2, pp. 180–192, 2008.

[44] M. Kovacs, “The children’s depression inventory (CDI),”Psychopharmacology Bulletin, vol. 21, pp. 995–998, 1981.

[45] C. F. Saylor, A. Spirito, and A. J. Finch Jr., “The children’sdepression inventory: a systematic evaluation of psychometricproperties,” Journal of Consulting and Clinical Psychology, vol.52, no. 6, pp. 955–967, 1984.

[46] M. R. Smucker, W. E. Craighead, L. Wilcoxon Craighead, andB. J. Green, “Normative and reliability data for the Children’sDepression Inventory,” Journal of Abnormal Child Psychology,vol. 14, no. 1, pp. 25–39, 1986.

[47] K. J. Preacher and A. F. Hayes, “Asymptotic and resamplingstrategies for assessing and comparing indirect effects inmultiple mediator models,” Behavior Research Methods, vol.40, no. 3, pp. 879–891, 2008.

[48] T. E. Moffitt, A. Caspi, H. Harrington et al., “Generalizedanxiety disorder and depression: childhood risk factors in abirth cohort followed to age 32,” Psychological Medicine, vol.37, no. 3, pp. 441–452, 2007.

[49] C. J. Lonigan, B. M. Phillips, and E. S. Hooe, “Relations ofpositive and negative affectivity to anxiety and depression inchildren: evidence from a latent variable longitudinal study,”Journal of Consulting and Clinical Psychology, vol. 71, no. 3,pp. 465–481, 2003.

[50] J. Roelofs, M. Huibers, F. Peeters, and A. Arntz, “Effectsof neuroticism on depression and anxiety: rumination as apossible mediator,” Personality and Individual Differences, vol.44, no. 3, pp. 576–586, 2008.

[51] J. Roelofs, M. Huibers, F. Peeters, A. Arntz, and J. vanOs, “Rumination and worrying as possible mediators in therelation between neuroticism and symptoms of depressionand anxiety in clinically depressed individuals,” BehaviourResearch and Therapy, vol. 46, no. 12, pp. 1283–1289, 2008.


[52] P. Muris, J. Roelofs, E. Rassin, I. Franken, and B. Mayer,“Mediating effects of rumination and worry on the linksbetween neuroticism, anxiety and depression,” Personality andIndividual Differences, vol. 39, no. 6, pp. 1105–1111, 2005.

[53] P. Muris, M. Fokke, and D. Kwik, “The ruminative responsestyle in adolescents: an examination of its specific link tosymptoms of depression,” Cognitive Therapy and Research, vol.33, no. 1, pp. 21–32, 2009.

[54] W. Kuyken, E. Watkins, E. Holden, and W. Cook, “Ruminationin adolescents at risk for depression,” Journal of AffectiveDisorders, vol. 96, no. 1-2, pp. 39–47, 2006.

[55] D. Lam, N. Smith, S. Checkley, F. Rijsdijk, and P. Sham, “Effectof neuroticism, response style and information processingon depression severity in a clinically depressed sample,”Psychological Medicine, vol. 33, no. 3, pp. 469–479, 2003.

[56] H. J. Eysenck and M. W. Eysenck, Personality and IndividualDifferences: A Natural Science Perspective, Plenum, New York,NY, USA, 1985.

[57] D. Derryberry and M. K. Rothbart, “Reactive and effortfulprocesses in the organization of temperament,” Developmentand Psychopathology, vol. 9, no. 4, pp. 633–652, 1997.

[58] TADS Team, “The treatment for adolescents with depressionstudy (TADS): long-term effectiveness and safety outcomes,”Archives of General Psychiatry, vol. 64, no. 10, pp. 1132–1144,2007.

[59] S. Jain, S. L. Shapiro, S. Swanick et al., “A randomizedcontrolled trial of mindfulness meditation versus relaxationtraining: effects on distress, positive states of mind, rumina-tion, and distraction,” Annals of Behavioral Medicine, vol. 33,no. 1, pp. 11–21, 2007.

[60] Z. Segal, J. M. G Williams, and J. D. Teasdale, Mindfulness-Based Cognitive Therapy for Depression: A New Approach forPreventing Relapse, Guilford Press, New York, NY, USA, 2002.

[61] J. Kabat-Zinn, “Mindfulness-based interventions in context:past, present, and future,” Clinical Psychology: Science andPractice, vol. 10, no. 2, pp. 144–156, 2003.

[62] P. R. Goldin and J. J. Gross, “Effects of mindfulness-basedstress reduction (MBSR) on emotion regulation in socialanxiety disorder,” Emotion, vol. 10, no. 1, pp. 83–91, 2010.

[63] A. M. Charbonneau and A. H. Mezulis, “A randomized trial ofa brief preventive intervention to reduce emotional reactivityto stress and depressive symptoms in the transition to college,”in Proceedings of the 31st Annual Meeting of the Society ofBehavioral Medicine, Seattle, Wash, USA, April 2010.

[64] R. A. Burwell and S. R. Shirk, “Subtypes of ruminationin adolescence: associations between brooding, reflection,depressive symptoms, and coping,” Journal of Clinical Childand Adolescent Psychology, vol. 36, no. 1, pp. 56–65, 2007.

[65] C. S. L. Lo, S. M. Y. Ho, and S. D. Hollon, “The effectsof rumination and negative cognitive styles on depression: amediation analysis,” Behaviour Research and Therapy, vol. 46,no. 4, pp. 487–495, 2008.

Submit your manuscripts athttp://www.hindawi.com

Stem CellsInternational

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014

Immunology ResearchHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinson’s Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttp://www.hindawi.com

Documents

RuminationMediatestheRelationshipbetween ...downloads.hindawi.com/journals/drt/2011/487873.pdf · birth (formerly named the Wisconsin Maternity Leave and Health Project; [39]). Participants